Aliso: A Journal of Systematic and Evolutionary Botany

Volume 14 | Issue 2 Article 2

1995 Wood and Bark Anatomy of Ranunculaceae (Including Hydrastis) and Glaucidiaceae Sherwin Carlquist Santa Barbara Botanic Garden

Follow this and additional works at: http://scholarship.claremont.edu/aliso Part of the Botany Commons

Recommended Citation Carlquist, Sherwin (1995) "Wood and Bark Anatomy of Ranunculaceae (Including Hydrastis) and Glaucidiaceae," Aliso: A Journal of Systematic and Evolutionary Botany: Vol. 14: Iss. 2, Article 2. Available at: http://scholarship.claremont.edu/aliso/vol14/iss2/2 Aliso, 14(2), pp. 65-84 © 1995, by The Rancho Santa Ana Botanic Garden, Claremont, CA 91711-3157

WOOD AND BARK ANATOMY OF RANUNCULACEAE (INCLUDING HYDRASTIS) AND GLAUCIDIACEAE

SHERWIN CARLQUIST Santa Barbara Botanic Garden 1212 Mission Canyon Road Santa Barbara, California 931051

ABSTRACT

Wood anatomy of 14 species of Clematis and one species each of Delphinium, Helleborus, Thal­ ictrum, and Xanthorhiza (Ranunculaceae) is compared to that of Glaucidium palma tum (Glaucidiaceae) and Hydrastis canadensis (Ranunculaceae, or Hydrastidaceae of some authors). Clematis wood has features typical of wood of vines and lianas: wide (earlywood) vessels, abundant axial parenchyma (earlywood, some species), high vessel density, low proportion of fibrous tissue in wood, wide rays composed of thin-walled cells, and abrupt origin of multiseriate rays. Superimposed on these features are expressions indicative of xeromorphy in the species of cold or dry areas: numerous narrow late­ wood vessels, presence of vasicentric tracheids, shorter vessel elements, and strongly marked growth rings. Wood of Xanthorhiza is like that of a (small) shrub. Wood of Delphinium, Helleborus, and Thalictrum is characteristic of herbs that become woodier: limited amounts of secondary xylem, par­ enchymatization of wood, partial conversion of ray areas to libriform fibers (partial raylessness). Wood of Ranunculaceae other than Clematis has numerous narrow vessels, probably an adaptation to cold more than drought. Glaucidium has occasional scalariform perforation plates like those almost uni­ versally present in Paeonia, but the two genera differ strongly in other wood features. Wood of Hydrastis (scalariform perforation plates occasional in primary xylem, rare in secondary xylem) ac­ cords with the idea that Hydrastis is a lineage that separated from the base of Ranunculaceae. Features that ally Ranunculaceae with other families of Ranunculiflorae include presence of wide multiseriate rays (without accompanying uniseriate rays), vessel restriction patterns, and storied structure. Wood­ iness in Ranunculaceae is probably secondary, at least to some extent; the woodiness of Xanthorhiza, thought by some to represent a primitive genus in the family, could be either primary or secondary. Key words: Berberidales, Glaucidium, Hydrastis, Paeonia, Papaverales, Ranunculaceae, Ranunculi- florae, storied wood structure, vessel restriction patterns.

INTRODUCTION omy of these families may help understand the inter­ relationships of these taxa. The present essay continues Phylogenetically, Ranunculaceae have been of in­ a series begun with studies of Lardizabalaceae (Carl­ terest because they have occupied a near-basal position quist 1984) and Papaveraceae (Carlquist and Zona in many systems. Most recently, the cladograms of 1988; Carlquist et al. 1994). Chase et al. (1993) follow this pattern on the basis of Despite the fact that Clematis contains about 250 rbcL data. Currently, Ranunculaceae are also of inter­ species of woody vines (Willis 1973), only a few spe­ est in relation to the paleoherb hypothesis (Taylor and cies have been investigated with respect to wood anat­ Hickey 1992). Anatomy of wood potentially offers omy. Accounts have been offered by Schmidt (1941), character expressions that illuminate these concepts. Metcalfe and Chalk (1950), Greguss (1959), Grosser Because Ranunculaceae contain a few woody genera (1977), Schweingruber (1978, 1992), and Sieber and as well as many herbaceous genera, it is a key family Kucera (1980). However, these accounts are incom­ in understanding primary versus secondary woodiness, plete, contradictory, and offer some inaccuracies. a chief concern of the paleoherb hypothesis. Wood anatomy of two species of Ranunculaceae in Ranunculiftorae (a superorder that equates with the genera other than Clematis has been briefly described orders Berberidales, Papaverales, Ranunculales, or by Schweingruber (1992), underlining the lack of Rhoeadales of earlier authors) consists of five main wood data for the family. In order to compare wood families: Berberidaceae, Lardizabalaceae, Menisper­ anatomy of Ranunculaceae to that of other Ranuncu­ maceae, Papaveraceae, and Ranunculaceae, as well as liftorae, a more nearly complete and accurate under­ some smaller satellite families. Studies on wood anat- standing of wood in the family than currently at hand is required. In this connection, information is needed 1 Address for correspondence with author: 4539 Via Huerto, Santa on wood of two genera traditionally included in Ran­ Barbara, CA 93110. This study was begun at the Rancho Santa Ana unculaceae (e.g., Prantl 1891): Hydrastis and Glauci­ Botanic Garden, 1500 N. College Ave., Claremont, CA 91711. dium. Recently excluded from Ranunculaceae as 66 ALISO monogenetic families, Glaucidium and Hydrastis have rhosa, and Glaucidium palmatum (see Table 1); por­ been the objects of research, chiefly on embryology tions of these three were preserved in 50% aqueous and floral anatomy (Tamura 1972; Tobe 1981; To be ethyl alcohol. Wood of genera other than Clematis, and Keating 1985). Keener (1993) has offered a sum­ Glaucidium, and Thalictrum were derived from spec­ mary of work to date on Hydrastis, and concludes that imens in the herbarium of the Rancho Santa Ana Bo­ it can be regarded as a subfamily within Ranuncula­ tanic Garden. Some of the wood samples of Clematis ceae. There is now general agreement that Paeonia were obtained from xylaria (see Table 1), and for these should be excluded from Ranunculaceae as Paeoni­ I am especially indebted to Dr. Regis Miller of the aceae and perhaps placed near the families of Ranun­ Forest Products Laboratory (MADw, SJRw) and to Dr. culiflorae (e.g., Thorne 1992), although not within the Roger Dechamps of the Musee Royale de l' Afrique group of families cited above as Ranunculiflorae. Centrale, Tervuren, Belgium (Tw). The woods of Wood anatomy may prove significant in comparing Clematis documented by my collection numbers (Ta­ Paeonia with Glaucidiaceae, Hydrastidaceae, and Ran­ ble 1) are all from cultivated specimens; herbarium unculaceae. specimens to document my collections are deposited Wood anatomy of Glaucidiaceae and Ranunculaceae at the Rancho Santa Ana Botanic Garden. I am grateful is worthy of examination with respect to habits and to the Municipal Botanic Garden, Dunedin, New Zea­ habitats. The woody vine Clematis ranges from upland land, and to the Rancho Santa Ana Botanic Garden for tropical to temperate alpine sites. The "woody herb" wood from cultivated specimens included in this study. (herb, with various degrees of secondary xylem pro­ The stem of Thalictrum polycarpum was collected duction in stems near the ground or beneath the ground from cultivated plants in the Santa Barbara Botanic surface) habits of some Ranunculaceae should be Garden. taken into account in studies of wood anatomy. Of the The localities of noncultivated specimens are as fol­ genera included in this study, only Xanthorhiza qual­ lows: C. dioscoriifolia, ruderal, New York City; C. ifies as a shrub or subshrub, an understory element in haenkeana, Ecuador; C. iringaensis, Shaba, Central moist temperate forests. Some species of the North Africa; C. ligusticifolia, dry riverbank near Yakima, Temperate genus Delphinium have a short, upright, Washington; C. paucifiora, Santa Ana Mountains, Cal­ woody caudex. Helleborus (Mediterranean region to ifornia; C. pickeringii, Fiji; Delphinium elatum, 83 the Caucasus Mountains) can be a subshrub with mod­ miles from Novosibirsk on highway to Bernaul, Si­ erately woody stems (H. foetidus L.), but most species, beria, Russia; Glaucidium palmatum, Mt. Shirano near such as H. viridis studied here, have rhizomes with Lake Chuzenji, Japan; Helleborus viridis, Montelupo, only a little secondary xylem. Hydrastis (two species Italy; Hydrastis canadensis, Neversink, Berks Co., of moist temperate forests: H. canadensis and H. je­ Pennsylvania; Xanthorhiza apiifolia, MacDowell Co., zoensis Sieb. & Zucc.) has succulent underground rhi­ North Carolina. The identity of the specimen cited here zomes. Glaucidium has a habit similar to that of Hy­ as C. haenkeana was determined with the aid of com­ drastis, but with more condensed rhizomes; the single ments from Dr. Frederick Essig and Dr. Nancy Mo­ species is confined to montane northern Japan (Prantl reno. 1891). Genera with minimal secondary xylem have Dried wood samples of both Clematis and the re­ been excluded from this study. maining genera (incorporating bark portions) provided Are Ranunculaceae primarily or secondarily woody, difficulties in sectioning because of wideness of ves­ or have both phyletic curricula occurred within the sels and (in bark) contrast between softness of phloem family? Wood anatomy examined in this context can and hardness of fibers. These problems were solved by cast light not only on this family but on others (such using a technique involving boiling the samples in wa­ as Paeoniaceae or Papaveraceae) as well. Because ter, softening them in ethylenediamine, embedding Ranunculaceae might be close to the "paleoherbs" hy­ them in paraffin, and sectioning them on a rotary mi­ pothesized to be primitive dicotyledons, this question crotome (Carlquist 1982). Most sections were stained has broader phyletic implications. in a safranin-fast green combination. Some of the par­ Because some of the samples studied here had intact affin sections were affixed to SEM aluminum stubs bark portions, some data on bark anatomy are reported. exactly as paraffin sections are mounted on glass These may be of interest both to systematics within slides; the paraffin was then dissolved with xylene so the family and to considerations of familial interrela­ that the sections could be studied by SEM. An lSI tionships. WB-6 scanning electron microscope was used. Mac­ erations were prepared by means of Jeffrey's Solution

MATERIALS AND METHODS and stained with safranin. Even though most wood samples had been dried rather than preserved in liquid, Wood samples were available in dried form with the preservation of starch and of primary walls proves ex­ exception of specimens of Clematis apiifolia, C. cir- cellent with this method. < 0 et'"' ~ tT1

-~- Table 1. Wood characteristics of Ranunculaceae and Glaucidiaceae. ez ~ I 2 3 4 5 6 7 8 9 10 II to Species Collection VG VD VM VL vw TL FN TW AP RW M tT1 :;o N Ranunculaceae Clematis alpina Mill. Carlquist 8051 00 13 783/1440 152 2-5 - VSEW+IVLW >20 1.4 C. apiifolia DC. Carlquist 8050 8.3 37 67/121 198 2-5 378 1.91 2-4 VSEW+IVLW 6.5 7.9 C. chrysocoma Franch. Carlquist 8056 >50 23 548/823 201 2-5 375 1.87 2-4 VSEW+IVLW >15 5.6 C. cirrhosa L. Carlquist 8053 >50 20 229/368 207 4-5 347 1.68 5 VSEW+IVLW 12.2 11.3 C. dioscoriifolia Lev. & Van. MADw-2756 >50 23 2611422 244 2-6 423 1.73 6 VAEW+VSLW >10 13.1 C. haenkeana Pres! SJRw-34198 8.3 37 671121 245 3-11 520 2.12 7 VAEW+VSLW 11.8 74.9 C. iringaensis Engler Tw-47313 1.7 97 18/36 225 2-6 501 2.23 5 VAEW+VALW 8.1 606.0 C. lasiantha Nutt. cult. RSABG >50 17 5711723 195 5-8 309 1.58 4 VSEW+IVLW 10.8 4.6 C. ligusticifolia Nutt. MADw-45101 >50 26 394/471 198 2-5 306 1.55 4 VSEW+IVLW 11.0 10.9 C. montana Buch-Ham. Carlquist 8049 >50 12 404/593 164 4 372 2.27 4 VSEW+IVLW 13.8 3.3 C. pauciflora Nutt. Wolf 2296 (RSAw) 00 20 524/654 195 2-5 364 1.67 5 VAEW+IVLW 8.1 6.0 C. pickeringii A. Gray SJRw-28381 11.3 37 121/205 253 2-7 482 1.91 5 VAEW+VSLW >10 45.7 C. tangutica Korsh. Carlquist 8107 >50 15 500/569 181 4 338 1.87 4 VSEW+IVLW 13.1 4.8 C. vitalba L. Carlquist 8109 >50 29 193/259 244 2-7 422 1.73 5 VAEW+IVLW 6.0 27.3 Delphinium elatum L. Smirnov 392 2.8 31 551112 135 2 312 2.31 2-7 VS,P 3.0 37.4 Helleborus viridis L. Sinclair 8755 4.1 13 434/964 142 5 306 1.73 1 IV >10 1.9 Hydrastis canadensis L. Brumbach 8592 3.7 13 176/386 160 1 250 1.56 1 p >10 5.4 Thalictrum polycarpum Wats. cult. SBBG 3.2 15 3311482 204 2 373 1.83 3 vs 4.4 6.3 Xanthorhiza apiifolia .LHer. Thorne 19304 4.5 20 198/285 167 1 342 2.05 3 Rare 10.3 11.7 Glaucidiaceae Glaucidium palmatum S. & Z. Carlquist 15726 4.5 34 1011149 159 2 313 1.97 5 P,IV >10 36.3

Key to columns: 1 (VG), mean number of vessels per group; 2 (VD), mean vessel diameter, j.Lm; 3 (VM), mean number of vessels per mm2-the first figure gives density of vessels in transections with rays included, the second the vessel density of fascicular areas only; 4 (VL), mean vessel element length, j.Lm; 5 (VW), vessel wall thickness, j.Lm; 6 (TL), mean length of imperforate tracheary elements (libriform fibers), j.Lm; 7 (FN), ratio, imperforate tracheary element length divided by vessel element length; 8 (TW), wall thickness of imperforate tracheary elements (libriform fibers), j.Lm; 9 (AP), axial parenchyma types: EW = earlywood, LW = latewood, IV = intervascular, P = pervasive, VA = vasicentric abundant, VS = vasicentric scanty; 10 (RW), mean width of ray at widest point, cells; 11 (M), Mesomorphy ratio (vessel diameter times vessel element length divided by vessels per mm2).

o-­ -...1 68 ALISO

Quantitative data are offered in Table I. Because C. vitalba as well as by Schweingruber (1992) for C. many vessels are flattened radially, tangentially, or oth­ campanulata, C. cirrhosa, and C. viticella L., ray-ad­ erwise, a compromise lumen diameter (intermediate jacent libriform fibers are evident, but these authors do between widest and narrowest measurement) was es­ not mention this phenomenon. Ray-adjacent fibers are timated for vessels noncircular in outline as a way of also present prior to origin of a multiseriate ray, as in representing conductive capacities more accurately. C. chrysocoma and C. vitalba. Libriform fibers are Exceptional care was taken to differentiate accurately also present in a mid-growth-ring position (compara­ among narrow vessels, vasicentric tracheids, and libri­ tively infrequent earlier or later in growth rings) in form fibers as seen in transection; the stains used aided fascicular areas of wood of C. dioscoriifolia (Fig. 3), cellular identification. Macerations permitted identifi­ C. cirrhosa, and C. iringaensis (Fig. 1). Libriform fi­ cation of vasicentric tracheids, which cannot be distin­ bers in positions abaxial to vessels were observed in guished from narrow vessel elements with certainty in C. apiifolia (Fig. 5) and C. ligusticifolia. Libriform longisections. Vasicentric tracheids have been over­ fibers are very scarce in Helleborus viridis, but are looked in Clematis, since there are no previous reports always present in a ray-adjacent position. In Delphin­ of them in the genus other than mine (Carlquist 1985). ium elatum (Fig. 16) and Glaucidium palmatum (Fig. These tracheids are called vasicentric rather than vas­ 22), libriform fibers are abundant in earlywood, chiefly cular for reasons given there (Carlquist 1985, p. 47). absent in latewood (some patchy or irregular distri­ Terms for vessel restriction patterns, ray types, and butions may be found, however), and tend to be more axial parenchyma types are based on Carlquist ( 1988). abundant in the earlier growth rings (which are wider) For other terms, the IAWA Committee on Nomencla­ than in the later ones, as shown for Glaucidium (Fig. ture ( 1964) has been used because of its accuracy and 22). In latewood of Delphinium elatum and Glauci­ applicability to purposes other than data computeriza­ dium palmatum, axial parenchyma tends to substitute tion. For each quantitative feature described, means for libriform fibers. The distribution of libriform fibers were derived from 25 measurements except for vessel in Hydrastis canadensis is similar (Fig. 26), but the wall thickness, libriform fiber wall thickness, and di­ fibers are in the middle of growth rings. mensions of lateral wall pits on vessels; for these fea­ If one looks at the above phenomenon from the as­ tures, typical conditions were selected for measure­ pect of vessel distribution rather than libriform fiber ment. distribution, a clear example of restriction of vessels to the central portions of fascicular xylem portions is ANATOMICAL DESCRIPTIONS seen in Xanthorhiza apiifolia (Fig. 20, 21). The assumption inherent in most descriptions of Growth Rings wood anatomy is that imperforate tracheary elements ("fibers") are distributed randomly. Likewise, descrip­ Growth rings are present in all Ranunculaceae stud­ tions of vessel distribution, other than those that relate ied, despite the fact that three species of Clematis are to growth rings, lead one to assume random distribu­ from tropical latitudes: C. haenkeana (Fig. 2, 15), C. tion of vessels. In Clematis, however, libriform fibers iringaensis (Fig. 1), and C. pickeringii (Fig. 7). Sea­ are commonly in a pattern that can be called ray-ad­ sonality in rainfall rather than temperature may influ­ jacent, along the edges of fascicular areas. Thus, ves­ ence growth ring formation in these species. Vasicen­ sels are infrequently adjacent to rays, and one could tric tracheids are absent in these tropical upland spe­ view this phenomenon as a vessel restriction pattern. cies. In the temperate species of Clematis, all of which Ray-adjacent distribution of libriform fibers is present have vasicentric tracheids, growth rings are highly pro­ in all of the Clematis species studied here except for nounced and vasicentric tracheids (intermixed with C. alpina, in which libriform fibers are absent (Fig. 4). narrow vessels) occur in latewood but at least a few Ray-adjacent fiber distributions are illustrated here for very wide vessels are in earlywood. All of the tem­ C. apiifolia (Fig. 5), C. chrysocoma (Fig. 6), C. cir­ perate species of Clematis, as well as Xanthorhiza api­ rhosa (Fig. 8), C. dioscoriifolia (Fig. 3), C. haenkeana ifolia and Thalictrum polycarpum, have growth rings (Fig. 2), C. iringaensis (Fig. 1), C. pickeringii (Fig. of this nature, which are illustrated here for C. alpina 7), and Xanthorhiza apiifolia (Fig. 20). In the illustra­ (Fig. 4), C. apiifolia (Fig. 5), C. chrysocoma (Fig. 6), tions of Greguss ( 1959) and Schweingruber ( 1992) for C. cirrhosa (Fig. 8), C. dioscoriifolia (Fig. 3), and

Fig. 1-4. Transections of Clematis wood.-I. C. iringaensis; ray-adjacent fibers flank the ray at left; axial parenchyma is abundant.- 2. C. haenkeana; strands of fibers occur at various points in the growth rings.-3. C. dioscoriifolia; cells of ray near center have thicker lignified walls below, but distal part of ray has thin primary walls on ray cells (some collapsed due to drying of wood sample).-4. C. alpina; five growth rings are shown. Collections given in Table 1. Arrows indicate zones of libriform fibers. (Scale for Fig. 1-4 above Fig. 1 [divisions = 10 f.Lm].) VOLUME 14, NUMBER 2 69 ______......

70 ALISO

Xanthorhiza apiifolia (Fig. 20). The abundance of nar­ ness of earlywood vessels and the relatively small row latewood vessels in these species accounts for the mean vessel diameter in the various species of Clem­ fact that they have much smaller mean vessel diameter atis is explained by the large number of narrow late­ (Table 1, column 2) as compared to C. haenkeana, C. wood vessels. In this respect, Clematis contrasts with iringaensis, and C. pickeringii. Growth rings are ex­ the other genera: although the other genera have mean tremely narrow in C. alpina (Fig. 4). Lack of vessels vessel diameters much like those in Clematis, they characterizes latewood in Helleborus foetidus accord­ have relatively uniform vessel diameter and lack no­ ing to Schweingruber (1992). Latewood of Delphinium tably wide vessel elements (Fig. 16, 18, 20, 22). The elatum (Fig. 16) and Glaucidium palmatum (Fig. 22) three tropical species of Clematis have relatively wide features axial parenchyma instead of libriform fibers. mean vessel diameters; these species have relatively If one views transections of fascicular areas of Clem­ few narrow latewood vessels compared to the temper­ atis alpina (Fig. 4) and C. pickeringii (Fig. 7), one ate species. sees apparent "indentations" in their outline. These · are due to presence of thin-walled axial parenchyma Vessel Density in latewood, adjacent to thin-walled ray cells. Mean vessel density is estimated in most studies of dicotyledon wood by averaging the vessel number as Vessel Grouping seen in random scans of transections. Such scanning Mean number of vessels per group (Table 1, column does not avoid rays, which do not comprise a large 1) ranges from 1.7 in C. iringaensis (Fig. 1) to an percentage of a transection in most dicotyledon woods. infinite number in C. alpina (Fig. 4), in which the en­ In the genera studied here, however, rays are large, and tirety of a fascicular secondary xylem area except for so vessel density has been calculated in Table 1, col­ scattered vasicentric tracheids and a few axial paren­ umn 3, in two ways: the first value is based on scans chyma cells consists of vessels. In most species of that include rays, the second figure is based on scans Clematis, the number of vessels per group averages of fascicular xylem only. A high value for vessel den­ more than 50 and cannot be counted exactly (Table 1, sity is evident in Clematis. The presence of large num­ column 1). Species with such extensive groupings in­ bers of narrow latewood vessels accounts for this. The clude C. apiifolia (Fig. 5), C. chrysocoma (Fig. 6), C. genera other than Clematis have relatively low vessel cirrhosa (Fig. 8), and C. dioscoriifolia (Fig. 3). Vessel density, in part because large patches of axial paren­ grouping is least in the species of tropical latitudes: C. chyma are present in some of them (Fig. 16, 22, 26). haenkeana (Fig. 2), C. iringaensis (Fig. 1), and C. In Helleborus viridis (Fig. 18), there are no large pickeringii (Fig. 7). Relatively low numbers of vessels patches of axial parenchyma but numerous narrow per group (2.8-4.5) characterize the genera studied vessels, so vessel density is relatively high. other than Clematis. All of these genera are woody herbs of moist forest, whereas Clematis is a woody Vessel Element Length vine of varied habitats. Correlations between quanti­ Mean vessel element length (Table 1, column 4) is tative vessel features and ecology are considered at the greater in the tropical species of Clematis (C. haen­ end of this paper. keana, C. iringaensis, C. pickeringii) than in the tem­ perate species. The species with the shortest vessel el­ Vessel Diameter ement length (152 1-1m) is C. alpina. Except for this Mean vessel diameter (Table 1, column 2) is not a species and C. montana, the genera other than Clem­ good indicator of the diversity of vessel diameter with­ atis mostly have shorter vessel elements than do Clem­ in a Clematis wood section. As the transections (Fig. atis species. However, these differences between 1-8) show, vessels exceeding 200 1-1m may be found Clematis and the other genera with respect to vessel in the early wood of most species. Vessels that wide element length is not very great. For one species, C. do not occur in C. alpina (Fig. 4), C. cirrhosa (Fig. pauciflora, vessel element length was calculated sep­ 8), C. montana, C. pauciflora, or C. tangutica, but the arately for wide earlywood vessel elements (vessel el­ widest vessels in these species are at least 100 1-1m in ements as wide as long or wider as seen in macera­ diameter. In macerations, very wide vessels are mostly tions) and narrow latewood vessel elements (a mean wider than long. The discrepancy between the wide- combining all vessel elements is shown in Table 1). In

Fig. 5-8. Transections of Clematis wood.-5. C. apiifolia; note marked contrast in vessel diameter between earlywood and latewood vessels.-6. C. chrysocoma; bands of fibers flank ray, left.-7. C. pickeringii; no growth rings are evident.-8. C. cirrhosa; ray cells resemble fibers in wall thickness; libriform fibers are adjacent to ray. Collections given in Table 1. Arrows indicate points at which ray cell walls change from lignified to unlignified. (Scales for Fig. 5-8 above Fig. 1.) VOLUME 14, NUMBER 2 71 72 ALISO

this species, mean length of wide vessel elements = alpina, C. apiifolia, C. dioscoriifolia, C. haenkeana, 162 J-Lm, mean length of narrow vessel elements = 217 C. tangutica, and C. vitalba. The vertical axis of pits J-Lm, and mean length of vasicentric tracheids = 218 is about 5 J-Lm in Clematis chrysocoma, C. cirrhosa, J-Lm. These data suggest that there is some intrusive C. iringaensis, C. lasiantha, C. ligusticifolia, C. pau­ growth in narrow vessel elements compared to wide ciflora, C. pickeringii, Delphinium elatum, Helleborus vessel elements as they mature. viridis, and Hydrastis canadensis. The vertical axis of pits is about 3 J-Lm in Thalictrum polycarpum and Xan­ Vessel Wall Thickness thorhiza apiifolia. Vessel wall sculpturing is prominent in most species In Clematis, vessel wall thickness (given in Table 1, of Clematis. In most species, the patterns include column 5) shows a range: less in narrow vessels, much grooves interconnecting pit apertures ("coalescent pit more in wider vessels. This can be seen at lower mag­ apertures") in wide vessels, combined with helical nifications in Fig. 1 and 2, but is much more evident thickenings-usually fine rather than coarse-in nar­ in Fig. 15. The genera other than Clematis have uni­ row vessels and vasicentric tracheids. This combina­ formly thin vessel walls except for Helleborus viridis, tion is shown for C. montana in Fig. 9, 10. The in which the vessel walls are notably thick (Fig. 18). grooves may be shallow, as at left in Fig. 9. The com­ bination of grooves in wide vessels plus fine helical Vessel Morphology thickenings in narrow vessels and vasicentric tracheids Simple perforation plates exclusively were observed was observed in C. apiifolia, C. chrysocoma, C. cir­ in all species except Glaucidium palmatum (Fig. 25) rhosa, C. dioscoriifolia, C. lasiantha, C. ligusticifolia, and Hydrastis canadensis. In H. canadensis, a single C. montana, C. pauciflora, Helleborus viridis (Fig. plate with one bar was seen, all other plates were sim­ 19), and Thalictrum polycarpum. Grooves in wide ves­ ple. Tobe and Keating (1985) figure scalariform plates sels only, and no sculpture in narrow vessels, were with 3-10 bars in H. canadensis, but these are metaxy­ observed in C. pickeringii. Grooves in both wider and lem perforation plates; they report that most perfora­ narrower vessels were observed in C. haenkeana and tion plates in the species are simple. Among the spe­ C. iringaensis. Coarse helical thickenings in both wide cies studied here, only in Glaucidium palmatum are and narrow vessels (slightly finer in the narrower ves­ scalariform perforation plates found in appreciable sels) were observed in C. vitalba (Fig. 11-13). The numbers, and even in my material of this species, only presence of helical thickenings in this species does not one of about 25 secondary xylem plates was observed exclude the presence of grooves interconnecting pit to have a single bar. The perforation plate shown in apertures; pairs of thickenings flanking grooves inter­ Fig. 25 was the only one observed to have more than connecting pit apertures could be seen on walls of wid­ one bar. A similar perforation plate of Glaucidium pal­ er vessels (Fig. 13). In Delphinium elatum, Glaucidium matum is figured by Tamura (1972), who terms it "re­ palmatum, Hydrastis canadensis, and Xanthorhiza api­ ticulate," although that plate would be termed scalar­ ifolia, no helical sculpture was observed on vessel iform by most workers. Tamura (1972) reports that walls. most perforation plates in Glaucidium palmatum are simple; thus, his observations are in accord with mine. Vasicentric Tracheids Lateral wall pitting of vessels (intervascular pitting) Vasicentric tracheids have been reported for one consists of alternate circular to oval pits in all species Clematis species earlier (Carlquist 1985), but they oc­ studies except Glaucidium palmatum. In that species cur in all of the species studied here except for the (Fig. 25), lateral wall pitting is either scalariform or three tropical species C. haenkeana, C. iringaensis, pseudoscalariform. Scalariform pits are defined as hav­ and C. pickeringii. In genera other than Clematis, vas­ ing a horizontal dimension extending the entire width icentric tracheids were observed only in Thalictrum. of a vessel wall face (vessels angular in transectional Vasicentric tracheids in Clematis and Thalictrum occur outline are assumed). Pseudoscalariform pits are ba­ mostly in latewood. Latewood distribution of these sically alternate but laterally widened, and their hori­ cells might lead one to designate them as vascular tra­ zontal axes do not coincide with vessel wall faces. The cheids, but that term is reserved for instances in which · vertical height of lateral wall pits is 8 J-Lm in Clematis tracheids (in a wood that also contains libriform fibers

Fig. 9-13. Scanning electron micrographs of vessel walls of Clematis from tangential sections.-9-10. C. montana.-9. Earlywood vessel wall has grooves interconnecting pit apertures.-10. Two latewood vessels have fine helical thickenings.-11-13. C. vitalba.-11. Wall of vessel of intermediate diameter; helical thickenings are coarse; grooves interconnect some of the pit apertures.-12. Two latewood vessels; helical thickenings are numerous.-13. Portion of wall of earlywood vessel; thickenings tend to occur as pairs of bands flanking pit apertures. Collections given in Table 1. (Scales shown at upper left in each figure [bars = 10 f.Lm].) VOLUME 14, NUMBER 2 73 ~------

74 ALISO

or fiber-tracheids) form terminal layers in latewood, Glaucidium (Fig. 22). Libriform fibers in Delphinium layers that do not include vessel elements to any ap­ elatum range in wall thickness from 2 to 7 ,_.,m; the preciable extent. In Clematis and Thalictrum, tracheids thicker-walled fibers occur in local patches. in latewood are intermixed with narrow vessels, so the term vasicentric must be applied. Sieber and Kucera Axial Parenchyma (1980) are correct in saying that Greguss ( 1959) errs Axial parenchyma in Clematis has been described in figuring helical thickening in a libriform fiber (with as apotracheal (Greguss 1959; Grosser 1977). Schmidt obviously simple pits) for C. vitalba. Possibly Greguss (1941), Metcalfe and Chalk (1950), and Schweingrub­ combined the helical thickenings seen in vessels or er (1978) consider the axial parenchyma in Clematis vasicentric tracheids with features of a libriform fiber. to be paratracheal. Sieber and Kucera (1980) and Schweingruber (1992) designate axial parenchyma in Libriform Fibers Clematis as both paratracheal and apotracheal. Sieber The imperforate tracheary element type (excepting and Kucera (1980) suspect that Greguss (1959) may vasicentric tracheids) of all Ranunculaceae studied is have mistaken narrow vessels for libriform fibers and the libriform fiber. Only simple pits were seen in these therefore declared axial parenchyma in contact with cells. Libriform fibers are lacking in my material of the alleged fibers to be apotracheal. In fact, when mac­ Clematis alpina (Fig. 4): the tissue in which vessels erations and sections were carefully examined, narrow are embedded consists of vasicentric tracheids and, to vessels were found to be very abundant in most spe­ a lesser extent, axial parenchyma. In Helleborus viri­ cies, and vasicentric tracheids were present in appre­ dis, libriform fibers are very scarce and are adjacent ciable numbers in all species of Clematis except the to ray areas. Libriform fibers in Clematis contain three tropical ones. Vasicentric tracheids may have starch (observed most clearly in wood samples that been mistaken for libriform fibers by some workers. were liquid-preserved: see Materials and Methods). Axial parenchyma in contact with vasicentric tracheids Septate libriform fibers were observed in C. picker­ probably should be classified as paratracheal, although ingii. Starch and septa are indications of prolonged I know of no worker who has considered this question. longevity of libriform fibers, and these features accord In any case, I failed to find any axial parenchyma that with Wolkinger's (1969) report of nucleated fibers in as individual cells or as part of larger parenchyma ag­ Clematis. gregations was not in contact with vessels. The large Mean length of libriform fibers (Table 1, column 6) masses of libriform fibers adjacent to rays (e.g., Fig. ranges from 250 to 520 ,_.,m. The longest libriform fi­ 1, 2, 3, 8) do not contain axial parenchyma. Thus, I bers occur in the tropical species: Clematis haenkeana, was unable to convince myself that any apotracheal C. iringaensis, and C. pickeringii. The genera other parenchyma was present in the specimens of Clematis than Clematis have relatively short libriform fibers. studied. The "FN ratio" (ratio of length of imperforate tra­ An unusual form of axial parenchyma is represented cheary elements-libriform fibers in the case of Ran­ in C. alpina (Fig. 4). In this species, the dark-staining unculaceae-to vessel elements) has been calculated fascicular xylem areas have an indented rather than for the species studied (Table 1, column 7). The values straight outline in contact with the rays. The apparent range from 1.55 to 2.31. The average figure for all indentations are, in fact, strands of thin-walled axial species studied is 1.87. Such a range is typical for a parenchyma adjacent to the rays in latewood areas. species with libriform fibers; lower F/V ratios (closer In Table 1, column 9, the axial parenchyma types to 1.0) would be expected for woods with tracheids or observed are shown. I have termed scattered axial pa­ fiber-tracheids. renchyma cells, distributed as single cells among nar­ Libriform fiber wall thickness is given in Table 1, row vessels (or vasicentric tracheids) "intervascular column 8. In Clematis, notably thick (5-7 1-1m) walls axial parenchyma" in accord with earlier definitions occur in C. cirrhosa (Fig. 8), C. dioscoriifolia (Fig. (Carlquist 1988). Where more than one axial paren­ 3), C. haenkeana (Fig. 2, 15), C. pauciflora, C. pick­ chyma cell, as seen in transection, is in contact with a eringii (Fig. 7), and C. vitalba. Libriform fibers were vessel, the term paratracheal scanty or vasicentric relatively thin walled in the other genera except for scanty (the latter from Kribs 1937) is used. In early-

Fig. 14-17. Wood sections of Ranunculaceae.-14-15. Clematis haenkeana.-!4. Tangential section; ray at left; storied libriforrn fibers (dark) and axial parenchyma (light, in strands of two cells) shown.-15. Transection; axial parenchyma is more abundant near earlywood vessel (below), less abundant near latewood vessels (center); transverse walls in axial parenchyma indicated by arrows.-16-17. Delphinium elatum.-!6. Transection; this portion is rayless, and has large, irregular patches of axial parenchyma.-17. Tangential section; the central third may be interpreted as a ray with markedly upright ray cells or a ray area converted to short libriform fibers. Collections given in Table I. (Scale for Fig. 14, 16, 17 above Fig. I; scale for Fig. 15 above Fig. 15 [divisions = 10 fLm].) VOLUME 14, NUMBER 2 75 76 ALISO wood of Clematis, the vessels are larger and therefore erogeneous Type III as uniseriate only, whereas rays each tends to be in contact with a few or even many in Clematis are wide multiseriate exclusively. Rays in axial parenchyma cells (Fig. I, 2, 15); where many all Ranunculaceae studied are rather wide (Table I, cells surround a vessel, the parenchyma may be termed column 10). vasicentric abundant. Axial parenchyma is very scarce Rays in most species of Clematis consist mostly of in Xanthorhiza apiifolia (Fig. 20). In Delphinium eta­ procumbent cells, but have one to two layers of upright tum (Fig. 16), Glaucidium palmatum (Fig. 22), and sheath cells. In some species such as C. alpina, C. Hydrastis canadensis (Fig. 26), large patches of axial haenkeana (Fig. 14), and C. iringaensis, three to four parenchyma are present in some areas of fascicular layers of sheath cells may be present, so that most of secondary xylem, whereas libriform fibers characterize the ray tissue consists of upright cells. In fact, rays of other zones of fascicular secondary xylem. Such ex­ Clematis do not fit any of the Kribs types. The rays tensive patches of axial parenchyma are termed "per­ of Clematis are best described as intermediate between vasive" in accordance with an earlier usage (Carlquist Homogeneous Type II (multiseriate, all cells procum­ 1995). In Helleborus viridis (Fig. 18), axial parenchy­ bent) and Paedomorphic Type II (multiseriate, upright ma is so common among vessels, rather than in group­ cells only). Paedomorphic Type II rays characterize C. ings around vessels or vessel groups, that the term pickeringii, and, judging from Schweingruber's (1992) "intervascular" appears appropriate. The terms inter­ illustrations, C. viticella. Curiously, Schweingruber vascular parenchyma and pervasive parenchyma were (1992) uses the term "Heterogeneous" for rays that not placed in usage in earlier decades because they are consist wholly of upright cells as well as for those in found chiefly in secondary xylem of relatively herba­ which both upright and procumbent cells are present. ceous dicotyledons, and wood of arboreal and shrubby Rays that consist wholly or predominantly of upright species has formed the basis for most works in wood cells should be termed Paedomorphic (Carlquist 1988). anatomy until recently. Paedomorphic Type II rays of this type were observed Sieber and Kucera (1980) report axial parenchyma in Delphinium elatum, Helleborus viridis, Glaucidium in strands of either two or three cells in Clematis vi­ palmatum (Fig. 24), Hydrastis canadensis, and Thal­ taiba. I agree with that designation, and certainly ictrum polycarpum. Rays of Xanthorhiza apiifolia strands of two are common in most wood samples of (Fig. 21) consist of upright plus some square cells, and Clematis, as shown in Fig. 14. Strands of two cells thus clearly qualify as Paedomorphic Type II, because exclusively were also observed in C. tangutica. How­ square cells are, by common consent, considered ever, careful examination of other Clematis woods re­ equivalent to upright cells morphologically. vealed wide fibriform cells with circular simple pits, In Clematis, new rays originate abruptly as wide thereby exactly like parenchyma in strands but formed multiseriate rays rather than as narrow rays that widen as undivided cells. Thus, strands of one to two cells later. This may be related to the scandent habit, be­ are reported here as constituting axial parenchyma of cause it also occurs in woody climbers in two other all species of Clematis except C. tangutica and C. vi­ families, neither closely related to Ranunculaceae: Cu­ taiba; strands of three or four cells, in addition to one curbitaceae (Carlquist 1992) and Aristolochiaceae or two, were observed in C. haenkeana wood. (Carlquist 1993). Another feature that may be related Axial parenchyma in strands of single undivided to the vining habit is the tendency in some species of cells characterizes Delphinium elatum (Fig. 17), Glau­ Clematis for rays to have at first cells that have lig­ cidium palmatum (Fig. 23), Helleborus viridis, and nified walls, then cells that are thin walled and nonlig­ Xanthorhiza apiifolia. In Hydrastis canadensis, strands nified. This, which is shown (some thin-walled ray of one to two cells were observed; in Thalictrum po­ cells collapsed, to be sure) for C. chrysocoma (Fig. 6, lycarpum, the strands consist of two to five cells. left), C. dioscoriifolia (Fig. 3, below), and C. picker­ ingii (Fig. 7, below), was also observed in C. montana. Rays Only thick-walled cells with lignified walls 1-4 f..lm Ray histology in Clematis has been characterized as thick were observed in C. apiifolia (Fig. 5), C. cirrho­ belonging to Heterogeneous Type III of Kribs (1935) sa (Fig. 8), C. haenkeana (Fig. 2), C. lasiantha, C. according to Sieber and Kucera (1980). This is clearly pauciflora, and C. vitalba. Bordered pits on tangential a mistake, because Kribs ( 1935, p. 551) defines Het- walls (and sometimes on other walls) were observed

Fig. 18-21. Wood sections of Ranunculaceae.-18-19. Helleborus viridis.-18. Transection; axial parenchyma is scattered among vessels rather than surrounding vessel groups.-19. Tangential section; inconspicuous helical thickenings are present on vessels (arrow).- 20-21. Xanthorrhiza apiifolia.-20. Transection; vessels are confined to a central position in the fascicular areas.-21. Tangential section; rays are composed of square to upright cells with secondary walls. Collections given in Table 1. (Fig. 18-19, scale above Fig. 15; Fig. 20, 21, scale above Fig. 1.) VOLUME 14, NUMBER 2 77 78 ALISO in C. znngaensis, C. pauciflora, and C. vitalba. Such upper half), Helleborus viridis (Fig. 18, lower half), bordered pits are often overlooked by workers because Hydrastis canadensis (Fig. 26), and Xanthorhiza api­ they view pits in face view, whereas the borders are ifolia (Fig. 20, 21). best seen in sectional view. In Clematis, starch grains occur in libriform fibers, The rays in Delphinium, Glaucidium, Helleborus, axial parenchyma, and ray cells. Likely any species of Hydrastis, Thalictrum, and Xanthorhiza are really only Clematis might be expected to contain starch in these extensions of primary rays. There can be histological wood cells, but starch might be expected to vary in changes in these rays during ontogeny, however. In presence according to season and according to method Delphinium, Glaucidium, and Thalictrum, some (but of wood sample preservation. Starch was observed not all) of the ray areas are soon converted to axial clearly in C. apiifolia, C. cirrhosa, C. iringaensis, and tissue; vertically elongate ray initials become almost Xanthorhiza apiifolia. No crystals were observed in the same length as fusiform cambial initials, and yield any wood cells in any of the Ranunculaceae studied. fusiform cells that qualify as libriform fibers rather than ray cells (Fig. 23). The central third of the tan­ Storied Structure gential section portion of Delphinium elatum shown in Fig. 17 is likely an area that began as a ray and was In older stems of Clematis, storying is evident. Sto­ replaced with wide libriform fibers not very different rying can be seen in all of the axial cell types: vessel from ray cells. This "partial raylessness" is a phenom­ elements, axial parenchyma, vasicentric tracheids, and enon not often encountered, and I am not able to cite even libriform fibers (Fig. 14). The storying of libri­ other examples. The relatively small width of rays in form fibers is less evident than that of the other cell Thalictrum polycarpum (Table 1, column 10) is due to types because of the elongation of the fibers, to about replacement of some ray areas with axial cells, much twice the length of vessel elements (Table 1, column as shown for Delphinium elatum in Fig. 17. The fig­ 7). The younger stems of Clematis I examined showed ures and descriptions of Schweingruber (1992) for less storying, in accord with what one finds elsewhere Thalictrum minus L. indicate only biseriate or unise­ in dicotyledons. No storying in axial xylem cells was riate rays. This condition is in contrast to other Ran­ observed in the genera of Ranunculaceae other than unculaceae-and Ranunculiflorae, in fact-that one is Clematis. tempted to question the determination of the wood Rays themselves are not storied in Clematis, but in sample. However, the transections Schweingruber C. alpina, storied ray cells may be seen in tangential (1992) illustrates for this species seem to show rapid sections. The rays in this species are markedly dilated conversion of ray tissue to fascicular tissue, achieving (noticeable especially in the low power transection a rayless condition, followed by initiation of narrow photograph by Schweingruber 1992). The storying of rays. This is a sequence that has been observed in ray cells in C. alpina seems likely the result of mul­ some other woods, such as Plantago arborescens Poir. tiplicative divisions in the ray initials of the cambium, (Carlquist 1970). Instances such as these are worthy but studies to confirm this were not undertaken. of further investigation. Quantitative data on ray width (Table 1, column 10) Bark Anatomy show that in both Clematis and the remaining genera, Only moderate amounts of phellem form on stems the rays are relatively little altered extensions of pri­ of Clematis (Fig. 27, 28). Periderms are successive mary rays, as indicated in the comments of Sieber and (Fig. 27). In the cortex, a few parenchyma cells may Kucera (1980) for C. vitalba. The relatively narrow enlarge into brachysclereids, as in C. pickeringii (Fig. ray widths given in the tabular data for Delphinium 28). Protophloem fibers form in all the species for elatum and Thalictrum polycarpum are related to the which bark was available (Fig. 28, upper right). In "partial raylessness" described above. most species of Clematis, annual increments of sec­ ondary phloem are completely encircled by a ring of Cell Contents fibers (Fig. 28). Other phloem increments are incom­ Large quantities of an unidentified yellowish mate­ pletely encircled, with fibers formed laterally and rial (often staining darkly in my material) were ob­ adaxially to each year's secondary xylem (Fig. 27, 29). served in vessels of Glaucidium palmatum (Fig. 22, In most species, phloem fibers do not form in central

Fig. 22-25. Wood sections of Glaucidium palmatum.-22. Transection; the portion shown is rayless, and axial parenchyma is mostly in latewood zones.-23. Tangential section showing a fascicular xylem portion; vessel elements are relatively short and have sca!ariform or preudoscalariform pitting.-24. Tangential section showing a wide ray that consists of thin-walled cells, many of which have collapsed.- 25. Vessels from a radial section showing scalariform and pseudoscalariform lateral wall pitting; the perforation plate (arrow) has three bars. (Scale for Fig. 22, 24 above Fig. I; scale for Fig. 23 above Fig. 15; scale for Fig. 25 above Fig. 25 [divisions = 10 fLm].) · VOLUME 14, NUMBER 2 79 ~------·..

80 ALISO

portions of phloem increments, but an exception is il­ mean vessel diameter, relatively few vessels per mm2 lustrated for C. apiifolia, in which strands of fibers of of transection, and relatively long vessel elements. various sizes occur (Fig. 27). Notably large diameter These three features have been combined as the Mes­ characterizes the sieve tube elements of Clematis (Fig. omorphy ratio in Table 1, column 11; the tropical spe­ 28, 29). cies show values from 45.7 to 606, higher than the In Xanthorhiza apiifolia, bark is like that illustrated values for any of the temperate species. Other features for Clematis iringaensis (Fig. 29), but with much characteristic of the tropical species are a relatively smaller increments of secondary phloem and with low number of vessels per group and an absence of much narrower sieve tube elements. The stems of vasicentric tracheids. Conditions alternative to those of Glaucidium studied were relatively old, and showed the tropical species characterize the temperate species. thick phellem, but with no sclerenchyma; secondary The Mesomorphy ratio is lowest for the montane tem­ phloem also lacks sclerenchyma in this genus. perate species C. alpina and C. montana. Only slightly less xeromorphic are the woods of Clematis species CONCLUSIONS from the Californian chaparral: C. lasiantha and C. paucifiora. Montane localities have physiological Habit and Ecology in Relation to Wood Anatomy drought due to freezing of soil; the chaparral species Differentiating between adaptations of wood that are experience soil dryness due to lack of rain during the related to ecology and those that correspond to habit six summer and autumn months. Vasicentric tracheids, is not possible in all instances. Certainly in Clematis present in the temperate species of Clematis, are char­ wood adaptations are primarily due to the climbing acteristic of woods from dry and cold areas (Carlquist habit; on those adaptations are superimposed those of 1985). climate, for Clematis ranges from tropical to temperate The genera studied other than Clematis are all from conditions. moist-temperate forest localities in which freezing of The scandent habit of Clematis is correlated with soil can occur in winter. Relatively narrow, short ves­ presence of limited numbers of wide thick-walled ves­ sel elements, numerous per mm2 of transection, as sels in earlywood and with a greater density of vessels found in these genera are characteristic of xeromorphic (in the wood as a whole) than found in all but dicot­ woods. The xeromorphy here is related to winter freez­ yledons with very xeromorphic wood. Libriform fibers ing, in all likelihood. These quantitative vessel fea­ are not randomly distributed throughout the wood, but tures, combined as the Mesomorphy ratio, range from rather mostly at the margins of fascicular areas, adja­ 1.9 to 37.4 (Table 1, column 11). Moist-temperate cent to rays. The rays are unusually wide and tall in shrubs would be expected to have values from 100 to Clematis (compared with those of dicotyledons at 500, and tropical wet-forest trees values above 1000, large), and are relatively infrequently subdivided dur­ so the subherbaceous genera other than Clematis have ing growth. Vasicentric tracheids are present in late­ a degree of xeromorphy equivalent to that of desert, wood and, less commonly, elsewhere in growth rings. alpine, or chaparral shrubs (Carlquist and Hoekman There is a tendency for rays to begin with cells that 1985). The low number of vessels per group in the have thick lignified walls, but over time these yield to genera other than Clematis is probably explained by thin-walled cells that have only primary walls (e.g., the relative abundance of parenchyma in these woods, Fig. 3, 6, 7). New rays are multiseriate from their or­ often occupying large patches ("pervasive axial paren­ igin, and are initiated abruptly (rather than as narrow chyma"). The rays are also wide in these genera. rays that gradually increase in width), as in Cucurbi­ Abundance of both axial and ray parenchyma in the taceae (Carlquist 1992) or Aristolochiaceae (Carlquist genera other than Clematis is likely related to their 1993). All of the above features tend to characterize subherbaceous habit, which features storage of pho­ wood of scandent taxa when those taxa are compared tosynthates in underground stems or basal portions of to nonscandent relatives (Carlquist 1991). stems during the winter months. Partial raylessness in Superimposed on the abovementioned wood fea­ Delphinium and Glaucidium is seen here as a mecha­ tures in Clematis are character expressions that relate nism for rapid production of more fibrous (and there­ to ecology. The three tropical species (C. haenkeana, fore mechanically strong) tissue, which can be of value C. iringaensis, and C. pickeringii) have relatively great in supporting tall upright stems that last only one year.

Fig. 26-29. Transections of stems of Ranunculaceae.-26. Secondary xylem of Hydrastis canadensis; axial parenchyma occurs in earlywood and latewood (arrows), libriform fibers are in an intermediate position in the growth rings.-27-29. Sections of bark of Clem­ atis.-27. C. apiifolia; strands of fibers occur within secondary phloem.-28. C. pickeringii; each increment of secondary phloem is sheathed in fibers; cortical sclereids, upper right and left.-29. C. iringaensis; fibers incompletely sheathe increments of secondary .phloem; note wide sieve tube elements. Collections given in Table I. (Scales for Fig. 26-29 above Fig. 1.) VOLUME 14, NUMBER 2 81 ~------

82 ALISO

Systematic Implications of Wood Anatomical Data phology and from epidermal micromorphology, show much the same pattern, although resolution of several Of the genera treated here, the one most commonly basal lineages in Ranunculaceae is poor in Hoot's considered not to belong to Ranunculaceae is Glau­ ( 1991) scheme. According to Hoot, Hydrastis, Xan­ cidium. Authors such as Tamura (1972) and Tobe thorhiza, and Coptis share the "T-type" small chro­ (1981), who have studied various aspects of this mosomes, in contrast to the "R-type" large chromo­ monotypic genus, make it the basis for a monogeneric somes elsewhere in the family. Wood of Hydrastis is family, Glaucidiaceae, and notice more resemblances not, however, like that of Xanthorhiza, although one between Glaucidiaceae and Paeoniaceae than between must keep in mind the differences in habit: Xantho­ Glaucidiaceae and Ranunculaceae. These ideas are re­ rhiza is a woody subshrub whereas Hydrastis is an flected in recent phylogenies, such as that of Thome herb with succulent, minimally woody prostrate stems. (1992), who assigns only Glaucidiaceae and Paeoni­ These differences could account for the differences in aceae to Paeoniales, but who assigns Ranunculaceae wood between these two genera. The presence in Hy­ and Hydrastidaceae to Berberidales. This treatment is drastis metaxylem of some scalariform perforation in agreement with the concepts of Tobe (1981), but plates, combined with presence in secondary xylem of Tamura (1972) assigns Paeoniaceae (and therefore also rare single-bar perforation plates (most plates are sim­ Glaucidiaceae) to Dilleniales, a treatment that is cer­ ple) is noteworthy. These xylary conditions are mar­ tainly current in some treatments (e.g., Cronquist ginally more primitive than those of Ranunculaceae 1988); Dilleniales is close to Theales or combined with s.s., in which scalariform perforation plates have been it by some systems. reported for primary xylem of two species of Clematis Is wood of Glaucidium like that of Paeonia? Glau­ and one of Ranunculus (Bierhorst and Zamora 1965); cidium does have scalariform perforation plates, but these authors report simple perforation plates exclu­ these are quite uncommon; plates with one bar are rare sively for primary xylem of one species each of Ane­ (about 1 in 25; the remainder of the plates are simple). monella (now Thalictrum) and Ranunculus. Zamora In Paeonia, by contrast, most perforation plates have (1966) listed 12 genera of Ranunculaceae in which 2-11 bars, and only 3.7% of the perforation plates are metaxylem contains " ... 2 of the following kinds of simple (Keefe and Moseley 1978). Lateral wall pitting perforation plates throughout the protoxylem-metaxy­ in Paeonia is transitional or alternate (very rarely sca­ lem transition: scalariform, transitional, simple." Sca­ lariform), whereas in Glaucidium it is scalariform or lariform perforation plates would be expected to per­ pseudoscalariform. The imperforate tracheary ele­ sist in primary xylem, as they do in Glaucidium and ments of Paeonia are tracheids, whereas they are al­ Hydrastis, in a phylad with simple perforation plates ways libriform fibers in Glaucidium. The axial paren­ in secondary xylem, if primary xylem is a "refugium" chyma of Paeonia is apotracheal and sparse, whereas for primitive vessel features, as Bailey (1944) hypoth­ in Glaucidium it is paratracheal and pervasive. Rays esized. The position of Hydrastis as a genus close to in Glaucidium are tall and wide and composed of cells Ranunculaceae and basally divergent from it appears with primary walls only, with no uniseriate rays pres­ justified; Keener (1993) makes a case for inclusion of ent. In Paeonia, rays are relatively narrow multiseriate Hydrastis in Ranunculaceae. The helical sculpturing of (2-8 cells wide) plus uniseriate; the ray cells have lig­ vessels of Hydrastis is a feature reminiscent of similar nified secondary walls. Thus, there are numerous con­ sculpture in vessels of Clematis and Thalictrum. Mo­ trasts between Paeonia and Glaucidium with respect lecular evidence may offer further evidence on wheth­ to wood anatomy. If Glaucidiaceae are indeed close to er Hydrastis should be segregated from Ranunculaceae Paeoniaceae, as embryological and other features have or retained within it. suggested to the authors cited earlier, the two families On the basis of data presented above for Ranuncu­ have diverged markedly with respect to wood anato­ laceae and Glaucidiaceae, as well as data in the liter­ my. ature for wood anatomy of families of Ranunculiftorae, Hydrastidaceae are recognized as a monogeneric three unusual wood features are worthy of mention as family separate from Ranunculaceae by some recent potential indicators of relationship. These are: presence authors (e.g., Tobe and Keating 1985; Thome 1992) of vessel restriction patterns (vessels confined to cen­ but not others (e.g., Cronquist 1988; Keener 1993). tral portions of fascicular areas); presence of wide Nowicke and Skvarla (1982) treated Hydrastidaceae as multiseriate rays exclusively (not accompanied by uni­ a member of Berberidales, which they are considering seriate rays); and presence of storied wood structure. as an order separate from Ranunculales. Tobe and Storied wood (and storying in fusiform cambial initials Keating (1985) conclude that Hydrastis diverged early and axial phloem cells), seen in Clematis, may also be from the basal stock of Ranunculaceae, and therefore seen in Lardizabalaceae (Carlquist 1984) and Papav­ represents a kind of relict element. The cladograms of eraceae (Carlquist and Zona 1988; Carlquist et al. Hoot (1991), which incorporate data from macromor- 1994), for example. Vessel restriction patterns, shown VOLUME 14, NUMBER 2 83 well in Clematis and Xanthorhiza, occur in Papaver­ ---. 1988. Comparative wood anatomy. Springer Verlag, Berlin aceae, as do the wide multiseriate rays (Carlquist and & Heidelberg. 436 p. ---. 1991. Anatomy of vine and Iiana stems: a review and Zona 1988, Carlquist et al. 1994). A summary paper synthesis, pp. 53-71. In F. E. Putz and H. A. Mooney [eds.], The is planned in order to map distribution of these key biology of vines. Cambridge University Press, Cambridge. features, as well as other wood features in Berberida­ ---. 1992. Wood anatomy of selected Cucurbitaceae and its ceae, Lardizabalaceae, Menispermaceae, Ranuncula­ relationship to habit and systematics. Nordic J. Bot. 12: 347-355. ceae and several other families claimed by particular ---. 1993. Wood and bark anatomy of Aristolochiaceae: sys­ tematic and habital correlations. lA WA J. 14: 341-357. authors to belong to Ranunculifiorae (e.g., Euptele­ ---. 1995. Wood anatomy of Caryophyllaceae: ecological, ha­ aceae). bital, systematic, and phylogenetic implications. Aliso 14: 1-17. The phylogenetic relationships of Ranunculaceae ---,AND D. A. HoEKMAN. 1985. Ecological wood anatomy of are important to the question of whether woodiness is the woody southern California flora.IAWA Bull., n.s., 6: 319-347. primary or secondary in Ranunculaceae-or, if both, ---,E. L. ScHNEIDER, AND R. B. MILLER. 1994. Wood and bark in which genera. At present, one should note that Ran­ anatomy of Argemone (Papaveraceae). lA WA J. 15: 247-255. ---, AND S. ZONA. 1988. Wood anatomy of Papaveraceae, with unculaceae have very short fusiform cambial initials comments on vessel restriction patterns. lA WA Bull., n.s., 9: 253- (as indicated by the length of vessel elements) as com­ 267. pared to Lardizabalaceae (Carlquist 1984) or Meni­ CHASE, M. W., D. E. SoLTIS, R. G. OLMSTEAD, D. MoRGAN, D. H. spermaceae (Metcalfe and Chalk 1950); both of these LES, B. D. MISHLER, M. R. DUVALL, R. A. PRICE, H. G. HILLS, Y.­ families are woody. Herbaceous phylads of dicotyle­ L. QIU, K. A. KRON, J. H. RETTIG, E. CONTI, J. D. PALMER, J. R. MANHART, K. J. SYTSMA, H. J. MICHAELS, W. J. KRESS, K. G. dons typically have short fusiform cambial initials in KAROL, W. D. CLARK, M. HEDREN, B. S. GAUT, R. K. JANSEN, K.­ comparison to those of their woody relatives. Short J. KiM, C. F. WIMPEE, J. F. SMITH, G. R. FURN!ER, S. H. STRAUSS, fusiform cambial initials in herbaceous dicotyledons Q.-Y. XIANG, G. M. PLUNKETT, P. S. SOLTIS, S.M. SWENSEN, S. E. may represent phyletic specialization of most herba­ WILLIAMS, P. A. GADEK, G. J. QUINN, L. E. EGU!ARTE, E. GOLEN­ ceous phylads rather than a feature related to the her­ BERG, G. H. LEARN, JR., S. W. GRAHAM, S. C. H. BARRETT, S. baceous habit per se. If any genus of Ranunculaceae DAYANADA, AND V. A. ALBERT. 1993. Phylogenetics of seed plants: an analysis of nucleotide sequences from the gene rbcL. were primitively woody, one would expect it to retain Ann. Missouri Bot. Gard. 80: 528-580. longer fusiform cambial initials. Whether or not Xan­ CRONQUIST, A. 1988. The evolution and classification of flowering thorhiza might be derived from woody ancestors is not plants. N. Y. Botanic Garden, Bronx, N.Y. 555 p. clear because cell size in Xanthorhiza wood is small GREGUSS, P. 1959. Holzanatomie der Europiiischen Laubhiilzer und in every respect. The rays of the specimen studied Striiucher. Akademiai Kiad6, Budapest. 330 p. were paedomorphic, suggesting the possibility of in­ GROSSER, D. 1977. Die HOlzer Mitteleuropas, Ein mikroskopischer Lehratlas. Springer Verlag, Berlin & Heidelberg. 208 p. terpreting the wood as secondarily woody. However, a HooT, S. B. 1991. Phylogeny of Ranunculaceae based on epidermal stem of maximal diameter for this genus was not avail­ microcharacters and macromorphology. Syst. Bot. 16: 741-755. able for the present study. An older stem is needed to IAWA COMMITTEE ON NOMENCLATURE. 1964. Multilingual glossary see whether the rays remain paedomorphic. Judging of terms used in wood anatomy. Verlagsbuchanstalt Konkordia, from the relatively basal position of Xanthorhiza in Winterthur, Switzerland. 186 p. Hoot's (1991) cladograms for Ranunculaceae, this is KEEFE, J. M., AND M. F. MOSELEY, JR. 1978. Wood anatomy and phylogeny of Paeonia section Moutan. J. Arnold Arb. 59: 274- the genus to which attention should be paid in assess­ 297. ing whether the family is primitively woody or her­ KEENER, C. S. 1993. A review of the classification of the genus baceous. Hydrastis (Ranunculaceae). Aliso 13: 551-558. KRIBs, D. A. 1935. Salient lines of structural specialization in the wood rays of dicotyledons. Bot. Gaz. 96: 547-557. LITERATURE CITED ---. 1937. Salient lines of structural specialization in the wood parenchyma of dicotyledons. 64: BAILEY, I. W. 1944. The development of vessels in angiosperms Bull. Torrey Bot. Club 177-186. L. and its significance in morphological research. Amer. J. Bot. 31: METCALFE, C. R., AND CHALK. 1950. Anatomy of the dicotyle­ 421-428. dons. Clarendon Press, Oxford. 1500 p. BIERHORST, D. W., AND P. M. ZAMORA. 1965. Primary xylem ele­ NOWICKE, J. W., AND J. J. SKVARLA. 1982. Pollen morphology and ments and element associations of angiosperms. Amer. J. Bot. 52: the relationships of Circaeaster, of Kingdonia, and of Sargento­ 657-710. doxa to the Ranunculales. Amer. J. Bot. 69: 990-998. CARLQUIST, S. 1970. Wood anatomy of insular species of Plantago PRANTL, K. 1891. Ranunculaceae. In A. Engler and K. Prantl [eds.], and the problem of raylessness. Bull. Torrey Bot. Club 97: 353- Die natiirlichen Pflanzenfamilien III(2): 43--66. Verlag Wilhelm 361. Engelmann, Leipzig. ---. 1982. The use of ethylenediamine in softening hard plant SCHMIDT, E. 1941. Mikrophotographischer Atlas der mittelleuro­ structures for paraffin sectioning. Stain Techno!. 57: 311-317. piiischen Holzer. Schriftenreihe des Reichsinstitutes fiir ausliin­ ---. 1984. Wood and stem anatomy of Lardizabalaceae, with dische and koloniale Forstwirtschaft no. 1. Verlag Neumann, Neu­ comments on the vining habit, ecology and systematics. Bot. J. damm. 112 p. Linnean Soc. 88: 257-277. SCHWEINGRUBER, F. H. 1978. Mikroskopische Holzanatomie. Edition ---. 1985. Vasicentric tracheids as a drought survival mecha­ Ziircher AG, Zug. 226 p. nism in the woody flora of southern California and similar regions; ---. 1992. Anatomie europiiischer Holzer. Verlag Paul Haupt, review of vasicentric tracheids. Aliso 11: 37-68. Bern & Stuttgart. 800 p. ;..~~·\ ------' ' 84 ALISO

SIEBER, M., AND L. J. KUCERA. 1980. On the stem anatomy of Clem­ ---,AND R. H. KEATING. 1985. The morphology and anatomy atis vitalba L. lA WA Bull., n.s., 1: 49-54. of Hydrastis (Ranunculales): systematic reevaluation of the genus. TAMURA, M. 1972. Morphology and phyletic relationships of the Bot. Mag. Tokyo 98: 291-316. Glaucidiaceae. Bot. Mag. Tokyo 85: 29-41. WILLIS, J. C. 1973. A dictionary of the flowering plants and ferns, TAYLOR, D. W., AND L. J. HICKEY. 1992. Phylogenetic evidence for the herbaceous origin of angiosperms. Pl. Syst. Evol. 180: 137- ed. 8., rev. by H. K. Airy-Shaw. Oxford University Press, Oxford. 156. WoLKINGER, F. 1969. Morphologie und systematische Verbreitung THORNE, R. F. 1992. Classification and geography of the flowering lebender Holzfasern bei Striiucher und Biiumen. I. Zur Morphol­ plants. Bot. Rev. 58: 225-348. ogie und Zytologie. Holzjorschung 23: 138-144. TOBE, H. 1981. Embryological studies in Glaucidium palmatum ZAMORA, P.M. 1966. Studies on the primary xylem elements in the Sieb. & Zucc. with a discussion on the taxonomy of the genus. Order Ranales (sensu lata): a systematic survey. Phillip. Agr. SO: Bot. Mag. Tokyo 94: 207-224. 389-623.